SWIFT: A Signature-enabled Wireless Infrastructure for Forensics,Tracking, and Locking of Electronic Systems
Abstract
The United States depends upon a range of critical systems and infrastructure to keep its society functioning, which places severe demands on the integrated circuits (ICs) and printed circuit boards (PCBs) comprising them. Unfortunately, todayÕs electronics are susceptible to a variety of threats, occurring both in the supply chain and in the field: counterfeiting, theft, reverse engineering, unauthorized access/use, and tampering. Such attacks create quality control issues and could lead to development of countermeasures that paralyze systems, hinder military objectives, and endanger lives. The Internet-of-Things (IoT) and advances in hardware security could result in game-changing supply chain management and system attestation capabilities for military agencies, law enforcement, and industry. In this PECASE project, we propose SWIFT (short for ÔSignature-enabled Wireless Infrastructure for Forensics, Tracking, and Locking of Electronic SystemsÕ) which augments traditional RFID systems by incorporating physical unclonable functions (PUFs), hardware obfuscation, and biometrics. In SWIFT, PUF-like signatures are captured from the RFID front end and electronic system (PCB). The signatures are designed to be sensitive to random variations and tampering in order to attest the authenticity and integrity of items being tracked in the supply chain. Through forensic/predictive analytics at the RFID back end (cloud), SWIFT can dynamically adapt supply chain transportation routes, deploy additional readers (such as drones) when sabotage/theft is detected, etc. In addition, system (PCB) level obfuscation, previously developed by the PI, is adopted as a built-in tamper response mechanism. If system signatures, operator biometrics, or other keys differ from their expected values in the field, the obfuscated system remains locked. These capabilities provided by SWIFT are highly relevant to the Army and DoD. Non-trivial challenges in SWIFT include generation of unique, reliable, and tamper evident signatures, protecting their privacy, and storage and mining of immense data. Addressing them requires expertise in modeling, design, and optimization across multiple domains Ð ICs, PCBs, biometrics, and distributed systems Ð which uniquely qualifies the PI to complete the proposed work. This project builds on his prior ARO Young Investigator Program award to resolve the above challenges and also investigate groundbreaking presentation attacks on biometric systems like SWIFT. There are five major tasks: (1) Variations occurring in the RFID antenna, tag, and power harvesting source will be modeled and optimized to extract unique and reliable signatures. Solutions based on standard PCB fabrication as well as promising, low-cost printable electronics will be explored; (2) Analog measurement modules and meshes will be added to the RFID front end and system to extract reliable signatures from traces on the PCB and protect them from tampering. Optimization will be performed to reduce the impacts of modifications to the PCB and compensate for environmental variations; (3) Protocols for securing the communication of signatures (RFID tag to reader and chips to obfuscation controller) will be developed and the security of different implementations will be analyzed. At this point, the entire RFID/PCB system will also be demonstrated using custom-made PCBs and COTS components; (4) We will analyze correlation in biometrics (fingerprint and face) resulting from genetic inheritance and develop novel presentation attacks that exploit them. To the best of our knowledge, this will be the first time such exploits have ever been investigated; and (5) We will consider novel back end and front end architectures that use compact data structures called Bloom filters. SWIFT storage and data mining costs, accuracy, and privacy will be modeled, simulated, analyzed, and optimized based on the proposed architectures.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Feb 14, 2019
- Source ID
- W911NF1910102
Entities
People
- Domenic Forte
Organizations
- Army Contracting Command
- United States Army
- University of Florida